Kernel Elements Research Articles

Intermediate-grained kernel elements pruning with structured sparsity

Neural network pruning provides a promising prospect for the deployment of neural networks on embedded or mobile devices with limited resources. Although current structured strategies are unconstrained by specific hardware architecture in the phase of forward inference, the decline in classification accuracy of structured methods is beyond the tolerance at the level of general pruning rate. This inspires us to develop a technique that satisfies high pruning rate with a small decline in accuracy and has the general nature of structured pruning. In this paper, we propose a new pruning method, namely KEP (Kernel Elements Pruning), to compress deep convolutional neural networks by exploring the significance of elements in each kernel plane and removing unimportant elements. In this method, we apply a controllable regularization penalty to constrain unimportant elements by adding a prior knowledge mask and obtain a compact model. In the calculation procedure of forward inference, we introduce a sparse convolution operation which is different from the sliding window to eliminate invalid zero calculations and verify the effectiveness of the operation for further deployment on FPGA. A massive variety of experiments demonstrate the effectiveness of KEP on two datasets: CIFAR-10 and ImageNet. Specially, with few indexes of non-zero weights introduced, KEP has a significant improvement over the latest structured methods in terms of parameter and float-point operation (FLOPs) reduction, and performs well on large datasets.

Low-Rank Kernel Tensor Learning for Incomplete Multi-View Clustering

Incomplete Multiple Kernel Clustering algorithms, which aim to learn a common latent representation from pre-constructed incomplete multiple kernels from the original data, followed by k-means for clustering. They have attracted intensive attention due to their high computational efficiency. However, our observation reveals that the imputation of these approaches for each kernel ignores the influence of other incomplete kernels. In light of this, we present a novel method called Low-Rank Kernel Tensor Learning for Incomplete Multiple Views Clustering (LRKT-IMVC) to address the above issue. Specifically, LRKT-IMVC first introduces the concept of kernel tensor to explore the inter-view correlations, and then the low-rank kernel tensor constraint is used to further capture the consistency information to impute missing kernel elements, thereby improving the quality of clustering. Moreover, we carefully design an alternative optimization method with promising convergence to solve the resulting optimization problem. The proposed method is compared with recent advances in experiments with different missing ratios on seven well-known datasets, demonstrating its effectiveness and the advantages of the proposed interpolation method.

Incomplete Multiple Kernel Alignment Maximization for Clustering.

Multiple kernel alignment (MKA) maximization criterion has been widely applied into multiple kernel clustering (MKC) and many variants have been recently developed. Though demonstrating superior clustering performance in various applications, it is observed that none of them can effectively handle incomplete MKC, where parts or all of the pre-specified base kernel matrices are incomplete. To address this issue, we propose to integrate the imputation of incomplete kernel matrices and MKA maximization for clustering into a unified learning framework. The clustering of MKA maximization guides the imputation of incomplete kernel elements, and the completed kernel matrices are in turn combined to conduct the subsequent MKC. These two procedures are alternately performed until convergence. By this way, the imputation and MKC processes are seamlessly connected, with the aim to achieve better clustering performance. Besides theoretically analyzing the clustering generalization error bound, we empirically evaluate the clustering performance on several multiple kernel learning (MKL) benchmark datasets, and the results indicate the superiority of our algorithm over existing state-of-the-art counterparts. Our codes and data are publicly available at https://xinwangliu.github.io/.

2Memristor-1Capacitor Integrated Temporal Kernel for High-Dimensional Data Mapping.

Compact but precise feature-extracting ability is core to processing complex computational tasks in neuromorphic hardware. Physical reservoir computing (RC) offers a robust framework to map temporal data into a high-dimensional space using the time dynamics of a material system, such as a volatile memristor. However, conventional physical RC systems have limited dynamics for the given material properties, restricting the methods to increase their dimensionality. This study proposes an integrated temporal kernel composed of a 2-memristor and 1-capacitor (2M1C) using a W/HfO2/TiN memristor and TiN/ZrO2/Al2O3/ZrO2/TiN capacitor to achieve higher dimensionality and tunable dynamics. The kernel elements are carefully designed and fabricated into an integrated array, of which performances are evaluated under diverse conditions. By optimizing the time dynamics of the 2M1C kernel, each memristor simultaneously extracts complementary information from input signals. The MNIST benchmark digit classification task achieves a high accuracy of 94.3% with a (196×10) single-layer network. Analog input mapping ability is tested with a Mackey-Glass time series prediction, and the system records a normalized root mean square error of 0.04 with a 20×1 readout network, the smallest readout network ever used for Mackey-Glass prediction in RC. These performances demonstrate its high potential for efficient temporal data analysis.

SEMAT Essence Game: A Fun Way to Learn about the SEMAT Initiative

Objective: we aim to propose the SEMAT Essence game, a game-based simulation of the big picture related to the SEMAT Essence kernel. SEMAT (Software Engineering Method and Theory) is an initiative for re-founding software engineering based on a solid theory, proven principles, and best practices. SEMAT has the Essence kernel of universal elements for any software engineering endeavor. The SEMAT Essence has been taught by using workshops and some games, but such strategies only cover some topics related to the kernel. Materials and methods: We use the SEMAT Essence kernel as a starting point and we select a game to be the template of our game. Then, we define the metaphor linked to our field of study and design the game by defining equivalences to the game elements. Results and discussion: We also show the results of practicing the game with a set of students from four universities by summarizing the answers to a survey. Conclusions: Results are promising since practitioners report some usage of the SEMAT Essence kernel elements after the game and they are satisfied with the fun factor, realism, and the current version of the game.

Analysis of the Aggregation Characteristics and Influencing Elements of Urban Catering Points in Small Scale: Methods and Results

Urban catering systems constitute an important subsystem of the complex urban system. They can reveal not only the impact of urban functional structure on the catering but also the behavioral patterns of individual catering points through the exploration of their small-scale aggregation characteristics and influencing elements, thus becoming an essential basis for urban functional planning. In this study, we analyze the aggregation characteristics of catering points in a particular study area using the probabilistic methods, with Beijing catering points as a sample. The analysis revealed a good power-law distribution characteristic of the catering points density at the small scale. Then, an aggregation effect analysis model and an agglomeration effect analysis model were established. Based on this, an empirical analysis of candidate agglomeration kernel elements was conducted. The results showed that the influence of candidate agglomeration kernel elements on catering points exhibited a categorical nature. Additionally, a good power-law attenuation relationship was uncovered between the density and distance of catering points, which ultimately revealed the mechanism of preferential attachment in the competition for catering point site selection. Using the results of the agglomeration analysis, a reasonable explanation was provided for the power-law distribution characteristic of the density of catering points, which achieved an organic connection between micro-analysis and macro-characteristic analysis. These findings could provide a reference for the analysis of aggregation characteristics of other urban commercial formats.

Asymmetric cost aggregation network for efficient stereo matching

AbstractCost aggregation is crucial to the accuracy of stereo matching. A reasonable cost aggregation algorithm should aggregate costs within homogeneous regions where pixels have the same or similar disparities. Otherwise, the estimated disparity map is prone to the well‐known edge‐fattening issue and the problem of losing fine structures. However, current state‐of‐the‐art convolutional neural networks (CNNs) mainly do cost aggregation with square‐kernel convolutional layers that learn to adjust their kernel elements to make the actual receptive fields of the aggregated costs adapt to homogeneous regions with various shapes. This is a mechanism that easily leads to the above issues due to the translation equivalence and content‐insensitivity properties of CNNs. To tackle these problems, a novel densely connected asymmetric convolution block (Dense‐ACB) based on asymmetric convolutions is proposed to explicitly construct receptive fields with various shapes, which effectively alleviates the issues caused by mismatching shapes of receptive fields and homogeneous regions. The proposed Dense‐ACB brings new insight to CNN‐based stereo matching networks. Based on the proposed cost aggregation method, an efficient and effective stereo matching network is built, which not only achieves competitive overall accuracy compared with state‐of‐the‐art models but also effectively alleviates the edge‐fattening problem and preserves fine structures.

Experimental Validation and Numerical Analysis of a High-Performance Blast Energy-Absorbing System for Building Structures

The paper presents a full-scale blast testing experimental campaign conducted on an energyabsorbing connector comprising thin-walled inversion tubes as kernel elements mounted in a façade protective panel. LS-DYNA finite element predictions of the global and local deformation/inversion of the panel/connectors compared reasonably well with the experimental observations. After validation, the numerical model was used to analyze the response of a simple idealized reinforced concrete structure under three blast-loading scenarios: the first two scenarios produce, approximately, the same impulse but are significantly different in terms of load duration and overpressures, and represent a far-field and a near-field scenario (1600 kg TNT at 20 m (i) and 150 kg TNT at 5 m (ii), respectively); the third scenario is more demanding, and consists in a half standoff distance of the second (150 kg TNT at 2.5 m (iii)). These numerical simulations allow to assess the effect of standoff distance and blast loading on the effectiveness of the protective system. One may conclude that the introduction of EACs strongly limits the forces imparted to the protected structure, reducing significantly the corresponding energy absorption demand. Comparing the energy absorbed by the structure in different scenarios, with and without the protective system (8 × ϕ64 × 2 mm), one can see that these reductions can reach, respectively 67%, 72% and 68% in the far-field, near-field and very near-field explosions.

CONSTRUCTION AND VERIFICATION OF A DIGITAL EQUALIZER MODEL

An approach to the development of an equalizer by building its mathematical model based on a microcontroller is proposed. All operations, including signal processing and equalizer kernel calculation, are performed by a single microcontroller. Thanks to the created mathematical model of the equalizer, the calculation of the kernel is reduced to multiple uses of relatively simple operations, which saves time and memory of the program. The equalizer provides satisfactory processing quality at a small filter order which is selected as a digital filter with final impulse response (FIR) because of its linear phase-frequency response, a guarantee of stability at the complex amplitude-frequency response, and also its associativity and linearity allowing it easily reproduce a complex amplitude-frequency response. Schematic implementation is based on parallel bandpass filters and a low-pass filter followed by adding filtered and amplified signals. It is the distributive property of the FIR filter that makes it possible to obtain a new kernel that includes all the amplified ranges by the sum of the corresponding kernel elements, instead of adding amplified ranges. The associativity and linearity of the FIR filter gives the opportunity to easily implement different types of filters on the basis of a low-pass filter, for the calculation of which the cardinal sine function is used together with the window function, which in combination gives qualitative frequency characteristics. The low-pass filter kernel and equalizer model are verified in the GNU Octave environment, which is an open-source analogue of Matlab. The model is checked by setting the frequency response of the test equalizer, and for individual filters the allowable width of the transition band and the maximum value of pulsation in the suppression band are set. The low-pass filter kernel is created with an arbitrary cutoff frequency, and the filter consists of 129 elements, which were multiplied by the Kaiser window with a value of parameter equal to 4.5. As a result of verification of the mathematical model in the GNU Octave environment, the width of the transition band and the maximum value of pulsation in the suppression band meet the specified conditions. The simulated equalizer kernel fully corresponds to the specified frequency response. Verification of the mathematical model confirmed its efficiency and compliance of the obtained characteristics of the equalizer with the specified requirements.

An Analytic Transform Kernel Derivation Method for Video Codecs

In the standardization of versatile video coding (VVC), discrete cosine transform (DCT)-2, discrete sine transform (DST)-7, and DCT-8 are regarded as the primary transform kernels. However, DST-4 and DCT-4 can also be considered as the transform kernels instead of using DST-7 and DCT-8 owing to their effectiveness in smaller resolution test sequences. To implement these different block size transform kernels, a considerable amount of memory has to be allocated. Moreover, memory consumption to store different block size transform kernels is regarded as a major issue in video coding standardization. To address this problem, a common sparse unified matrix concept is introduced in this study, where any block size transform kernel matrix can be obtained after some mathematical operations. The proposed common sparse unified matrix saves approximately 80% of the static memory by storing only a few transform kernel elements for DCT-2, DST-7, and DCT-8. Full-required transform kernels are derived using the stored transform kernels and generated unit-element matrices and a permutation matrix. The static memory required is only for 1648 elements instead of 8180 elements, each with 8-bit precision. The defined common sparse unified matrix is composed of two parts: a unified DST-3 matrix and a grouped DST-7 matrix. The unified DST-3 matrix is used to derive different points of DCT-2 transform kernels, and the grouped DST-7 matrix is used to derive different points of DST-7 and DCT-8 transform kernels. The new technique of grouping concept is introduced, which shows the relationship between different rows of DST-7 transform kernels with various block sizes. The proposed grouping concept supports the fast algorithm of DST-7 by implementing the proposed method of the “one group one feature” principle. The simulation was conducted using the VTM-3.0 reference software under common test conditions. The simulation result of the all intra (AI) configuration is Y = 0.00%, U = −0.02%, V = 0.00% with an encoding time of 100%, and a decoding time of 100%. Similarly, the simulation results of random access (RA) configuration are Y = −0.01%, U = 0.09%, V = 0.06%, and the encoding and decoding times are 101% and 100%, respectively. The simulation result of the low delay B (LDB) configuration is Y = 0.01%, U = 0.08%, and V = −0.27%, for encoding and decoding times of 101% and 100%, respectively.

High energy flash X‐ray image restoration using region extrema and kernel optimization

The quality of high energy flash X-ray images is crucial to the high-precision diagnosis of object density. High energy flash X-ray radiography is susceptible to the system blur, which usually causes the poor quality of static images. In response to this, a novel restoration algorithm using region extrema and kernel optimization (REKO) is presented. Based on the observation that the region extrema distribution of blurred high energy flash X-ray images deviates from opposite ends of image grey domain, the sparseness-inducing prior for regularizing image region extrema is applied to construct the restoration model. Considering the sparse characteristics of blur kernels, the sparseness-inducing regularization is incorporated to constrain blur kernels in the restoration model. The non-convex and non-linear objective function is gradually minimized through energy alternating minimization and dually linear approximation. Furthermore, a continuity enforced kernel optimization algorithm is proposed to estimate more accurate blur kernels. The discontinuous kernel elements are suppressed by extracting the main structure of blur kernels and constructing kernel continuity function in cross windows. Experimental results demonstrate that our algorithm can more accurately estimate blur kernels and achieve restoration results with sharper edges on high energy flash X-ray images.

An innovative seismic-resilient bridge with shape memory alloy-washer-based footing rocking RC piers

This study proposes a new footing rocking pier which employs shape memory alloy (SMA) washers as its kernel elements. Apart from providing self-centering capability and energy dissipation, the SMA washers enable a “limited rocking” behavior for the pier, helping enhance the seismic resilience of the bridges. The study commences with quasi-static tests on SMA washer groups with various stack patterns. Subsequently, a 1/4 scale SMA-washer-based rocking pier specimen is tested. Following the experimental study, a bridge system adopting the SMA-washer-based rocking pier is carefully designed. An iterative seismic design approach is proposed for the rocking pier, followed by a system level analysis on the entire bridge system. A suite of ground motions corresponding to two intensity levels are considered. The performance of the innovative bridge is also compared with a conventional bridge with fixed-base piers. The results reveal that the new system exhibits superior performance to the conventional bridge in terms of pier damage control, which is achieved with acceptable sacrifice of other performance indices such as bearing displacement and peak drift ratio. A further parametric study suggests that the response of the bridge with the proposed pier is related to the load resistance and deformability of the SMA washers.

On SU(3)F positive-parity octet and decuplet baryons

A continuum approach to the three valence-quark bound-state problem in quantum field theory, employing parametrisations of the necessary kernel elements, is used to compute the spectrum and Poincarö- covariant wave functions for all flavour-SU(3) octet and decuplet baryons and their first positive-parity ex citations. Such analyses predict the existence of nonpointlike, dynamical quark-quark (diquark) correlations within all baryons; and a uniformly sound description of the systems studied is obtained by retaining flavour- antitriplet-scalar and flavour-sextet-pseudovector diquarks. The analysis predicts the existence of positive- parity excitations of the 𝚵, 𝚵*, Ω baryons, with masses, respectively (in GeV): 1.84(08), 1.89(04), 2.05(02). These states have not yet been empirically identified. This body of analysis suggests that the expression of emergent mass generation is the same in all u, d, s baryons and, notably, that dynamical quark-quark correla tions play an essential role in the structure of each one. It also provides the basis for developing an array of predictions that can be tested in new generation experiments.

Robust Motion Blur Kernel Estimation by Kernel Continuity Prior

The accurate kernel estimation is key to the blind motion deblurring. Many previous methods depend on the image regularization to recover strong edges in the observed image for kernel estimation. However, the estimated kernel will be degraded when recovered strong edges are less accurate, especially in images full of small-scale edges. Different from previous methods, we focus on the kernel regularization. Inspired by the fact that the blur kernel is highly related to the continuous camera motion trajectory during the image capturing, we propose to encourage the continuity of the kernel through a kernel prior. The proposed prior measures the continuity of each element in the kernel and generates a continuity map. By encouraging the sparsity of the map using L 0 norm, discontinuous kernel elements are suppressed. Since the model with the proposed prior is non-convex and non-linear, an approximation method is proposed to minimize the cost function efficiently. Numerous experimental results show that our method outperforms state-of-the-art methods on both the normal and challenging cases. Moreover, the proposed prior is able to further improve the performance of existing MAP-based methods.

Spectrum and structure of octet and decuplet baryons and their positive-parity excitations

A continuum approach to the three valence-quark bound-state problem in quantum field theory, employing parametrisations of the necessary kernel elements, is used to compute the spectrum and Poincar\'e-covariant wave functions for all flavour-$SU(3)$ octet and decuplet baryons and their first positive-parity excitations. Such analyses predict the existence of nonpointlike, dynamical quark-quark (diquark) correlations within all baryons; and a uniformly sound description of the systems studied is obtained by retaining flavour-antitriplet--scalar and flavour-sextet--pseudovector diquarks. Thus constituted, the rest-frame wave function of every system studied is primarily $S$-wave in character; and the first positive-parity excitation of each octet or decuplet baryon exhibits the characteristics of a radial excitation. Importantly, every ground-state octet and decuplet baryon possesses a radial excitation. Hence, the analysis predicts the existence of positive-parity excitations of the $\Xi$, $\Xi^\ast$, $\Omega$ baryons, with masses, respectively (in GeV): 1.84(08), 1.89(04), 2.05(02). These states have not yet been empirically identified. This body of analysis suggests that the expression of emergent mass generation is the same in all $u$, $d$, $s$ baryons and, notably, that dynamical quark-quark correlations play an essential role in the structure of each one. It also provides the basis for developing an array of predictions that can be tested in new generation experiments.

Short-Term Reliability Assessment of UHVdc Systems Based on State Aggregation With SMP

The short-term reliability of the ultra-high-voltage dc (UHVdc) systems has enormous influence on the security of the whole power system. The frequency and duration (F&D) algorithm based on the Markov process (MP) will reduce the accuracy of the short-term reliability indices, due to the exponentially distributed parameters and the strict premises for the state aggregation and combination. In this paper, the short-term aggregation algorithm based on the semi-MP (SMP) is improved. The aggregated transfer rates are obtained by the definition of the semi-Markov kernel elements and the invariability of the transfer probabilities before and after aggregation. The algorithm is applied to the series structure system, which is common in UHVdc. The aggregated kernel matrix and instantaneous transfer probability matrix are derived from those before aggregation directly. A new method is proposed to describe the combined transfer rates by matrices combination and deleting the nonexistent states and transitions, which requires less calculation effort. The Erlang and Weibull distributions are first used to fit the bell-shaped probability density functions (PDFs) of the repair times of the equipments in the UHVdc. The numerical examples show that, compared with the SMP algorithm, the traditional F&D algorithm overestimates the instantaneous equivalent outage hours of the UHVdc. The error is originated from that the F&D algorithm regards the aggregated system as a MP while it is non-Markovian. To show the effectiveness of the bell-shaped repair time, the Erlang and Weibull distributions are compared with the exponential one by the reliability indices. The differences are result from the different PDFs of these distributions.

Integral Accuracy and the Stability of Two Methods for the Solution of Time-Domain Integral Equations for Scattering From Perfect Conductors

The stability of time-domain (TD) integral equation (TDIE) approaches to the computation of electromagnetic scattering is profoundly affected by the accuracy of the underlying numerical integration methods used for computation of the kernel elements. In most publications, the accuracy of the quadratures used to compute kernel elements is never measured, and higher order computation is assumed to deliver high accuracy. In this communication, we examine the complicated relationship between the actual accuracy of kernel element computation and the resulting stability of integral equations. The numerical results show that stability may be improved for higher integral accuracy. Although integral accuracy is not always improved by higher order integration rules, more careful integration (as delivered by adaptive integration methods) is often helpful. The numerical results for a range of problems demonstrate these contentions.

Manufacturing and performance of a novel self‐centring damper with shape memory alloy ring springs for seismic resilience

This paper discusses the manufacturing process and mechanical performance of a novel self-centring damper equipped with a group of kernel elements, namely, shape memory alloy (SMA) ring springs. The study commences with an experimental investigation on individual SMA ring spring specimens, enabling a fundamental understanding of their working principle and mechanical performance. Some technical issues related to manufacturing process, low-temperature heat treatment (annealing), quenching, and cyclic training are also discussed. Subsequently, the fabrication steps and working mechanism of the dampers incorporating the SMA ring springs are elaborated, and a further experimental study on a large-scale prototype damper specimen is carried out. The test results show that the damper specimen exhibits satisfactory mechanical behaviour with a stable flag-shaped load–displacement hysteretic response. A high initial stiffness and an adequate level of “yield” resistance are exhibited due to a relatively high preload applied to the SMA ring spring group. Excellent self-centring capability with no residual displacement is observed, and satisfactory energy dissipation with an equivalent viscous damping ratio of up to 18.5% is shown. No failure to any component of the damper is observed by the end of the test. The proposed damper is expected to resist strong earthquakes and offer self-centring driving action effectively for engineering structures without the need for repair.

Multi-PSF fusion in image restoration of range-gated systems

For the task of image restoration, an accurate estimation of degrading PSF/kernel is the premise of recovering a visually superior image. The imaging process of range-gated imaging system in atmosphere associates with lots of factors, such as back scattering, background radiation, diffraction limit and the vibration of the platform. On one hand, due to the difficulty of constructing models for all factors, the kernels from physical-model based methods are not strictly accurate and practical. On the other hand, there are few strong edges in images, which brings significant errors to most of image-feature-based methods. Since different methods focus on different formation factors of the kernel, their results often complement each other. Therefore, we propose an approach which combines physical model with image features. With an fusion strategy using GCRF (Gaussian Conditional Random Fields) framework, we get a final kernel which is closer to the actual one. Aiming at the problem that ground-truth image is difficult to obtain, we then propose a semi data-driven fusion method in which different data sets are used to train fusion parameters. Finally, a semi blind restoration strategy based on EM (Expectation Maximization) and RL (Richardson-Lucy) algorithm is proposed. Our methods not only models how the lasers transfer in the atmosphere and imaging in the ICCD (Intensified CCD) plane, but also quantifies other unknown degraded factors using image-based methods, revealing how multiple kernel elements interact with each other. The experimental results demonstrate that our method achieves better performance than state-of-the-art restoration approaches.

Fischer’s Fate with Fatalism

John Martin Fischer’s core project in Our Fate (2016) is to develop and defend Pike-style arguments for theological incompatibilism, i. e., for the view that divine omniscience is incompatible with human free will. Against Ockhamist attacks on such arguments, Fischer maintains that divine forebeliefs constitute so-called hard facts about the times at which they occur, or at least facts with hard ‘kernel elements’. I reconstruct Fischer’s argument and outline its structural analogies with an argument for logical fatalism. I then point out some of the costs of Fischer’s reasoning that come into focus once we notice that the set of hard facts is closed under entailment.